w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/watres

Graphene oxide based ultrafiltration membranes

for photocatalytic degradation of organic pollutants
in salty water

Luisa M. Pastrana-Martı́nez, Sergio Morales-Torres, Jos
e L. Figueiredo,

n M.T. Silva*
Joaquim L. Faria, Adria
LCM e Laboratory of Catalysis and Materials e Associate Laboratory LSRE-LCM, Faculdade de Engenharia,
Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal

article info abstract

Article history: Flat sheet ultrafiltration (UF) membranes with photocatalytic properties were prepared
Received 16 October 2014 with lab-made TiO2 and graphene oxide-TiO2 (GOT), and also with a reference TiO2 pho-
Received in revised form tocatalyst from Evonik (P25). These membranes were tested in continuous operation mode
25 February 2015 for the degradation and mineralization of a pharmaceutical compound, diphenhydramine
Accepted 15 March 2015 (DP), and an organic dye, methyl orange (MO), under both near-UV/Vis and visible light
Available online 24 March 2015 irradiation. The effect of NaCl was investigated considering simulated brackish water (NaCl
0.5 g L
1) and simulated seawater (NaCl 35 g L
1). The results indicated that the membranes
Keywords: prepared with the GOT composite (M-GOT) exhibited the highest photocatalytic activity,
Graphene oxide outperforming those prepared with bare TiO2 (M
TiO2) and P25 (M-P25), both inactive
Photocatalytic membranes under visible light illumination. The best performance of M-GOT may be due to the lower
Filtration band-gap energy (2.9 eV) of GOT. In general, the permeate flux was also higher for M-GOT
Water purification probably due to a combined effect of its highest photocatalytic activity, highest hydro-
Salty water philicity (contact angles of 11 , 17 and 18 for M-GOT, M
TiO2 and M-P25, respectively) and
Anti-fouling higher porosity (71%). The presence of NaCl had a detrimental effect on the efficiency of the
membranes, since chloride anions can act as hole and hydroxyl radical scavengers, but it
did not affect the catalytic stability of these membranes. A hierarchically ordered mem-
brane was also prepared by intercalating a freestanding GO membrane in the structure of
the M-GOT membrane (M-GO/GOT). The results showed considerably higher pollutant
removal in darkness and good photocatalytic activity under near-UV/Vis and visible light
irradiation in continuous mode experiments.
© 2015 Elsevier Ltd. All rights reserved.

well as in less developed regions. In this context, efficient

1. Introduction water purification technologies with low energy consump-
tions are needed (Shannon et al., 2008). Membranes play a key
The scarcity of clean water and increasing environmental role in water purification, seawater and brackish water desa-
pollution are critical issues in large and industrialized cities as lination as well as in wastewater reclamation and reuse

* Corresponding author. Tel.: þ351 22 0414908.

E-mail address: adrian@fe.up.pt (A.M.T. Silva).
http://dx.doi.org/10.1016/j.watres.2015.03.014
0043-1354/© 2015 Elsevier Ltd. All rights reserved.
180 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
(Pendergast and Hoek, 2011; Peters, 2010). The preparation of
appropriate membranes is an important step, and different
materials, such as polymers, ceramics and carbon nanotubes,
have been successfully employed (Baek et al., 2014; Das et al.,
2014; Ulbricht, 2006; Xu et al., 2013b). Fouling is one of the
major problems affecting the performance of these
membranes.
In recent years, membranes prepared with photocatalytic
nanoparticles have attracted great attention due to their su-
perior characteristics (e.g., anti-fouling and photocatalytic
properties) when compared to conventional membranes (Kim
and Van der Bruggen, 2010). Titanium dioxide (TiO2) is one of
the most widely used photocatalysts due to its low cost,
chemical and thermal stability and excellent photoactivity
(Chen and Mao, 2007). Up to date, various TiO2ebased photo-
catalytic membranes have been studied (Albu et al., 2007; Pan
et al., 2008; Zhang et al., 2006, 2014b). However, the suppres-
sion of the recombination of photo-generated charge carriers,
as well as the effective utilization of visible light, are some of
the main challenges before these membranes become
economically feasible.
Graphene and its derivatives, such as graphene oxide (GO),
have attracted huge attention in photocatalytic applications
(Tu et al., 2013). GO is a material that can be easily produced by
chemical oxidation and exfoliation of graphite. In addition, it
can be easily manipulated and its oxygen-containing func-
tional groups (hydroxyl and epoxy groups on the basal planes,
and carboxyl and carbonyl groups on the edges) facilitate the
interaction of GO sheets with a wide variety of organic and
inorganic materials (Dreyer et al., 2010; Pastrana-Martı́nez
et al., 2014). GO-TiO2 composites are efficient photocatalysts
under both near-UV/Vis and visible light irradiation, over-
coming one of the main limitations of bare TiO2 (Amalraj
Appavoo et al., 2014; Fan et al., 2011; Huang et al., 2014; Long
et al., 2013; Pastrana-Martı́nez et al., 2013a; 2014; 2012). How-
ever, these materials are usually employed as suspended
particles (slurries) in batch reactors and, thus, a second step is
required for catalyst separation from the treated water,
limiting its recovery and reuse.
The immobilization of the photocatalyst into/onto filtra-
tion membranes/fibres overcomes this problem, but only a
few works have been published on this topic (Athanasekou
et al., 2014; Gao et al., 2013, 2014; Pastrana-Martı́nez et al.,
2013b). A GO-TiO2 microsphere hierarchical membrane was
developed by assembling the photocatalyst on the surface of a
cellulose acetate filtration membrane, which showed the
multifunctionality of water filtration and photodegradation of
acid orange 7 and rhodamine B (Gao et al., 2013). A GO-TiO2
photocatalytic membrane was synthesized by simple layer-
by-layer deposition of TiO2 and GO on a polysulfone mem-
brane, and the increase in the membrane flux was attributed
to the photo-enhanced hydrophilicity and simultaneous
degradation of the methylene blue model pollutant (Gao et al.,
2014). Recently, we immobilized a highly active GO-TiO2
composite into alginate hollow fibres by a dry/wet spinning
process, and considerable high photocatalytic activity and
stability for degradation of diphenhydramine were observed
in consecutive lightedark cycles of continuous operation
(Pastrana-Martı́nez et al., 2013b). This composite was also
immobilized onto ceramic monoliths via dip-coating and
tested as a hybrid photocatalysis/ultrafiltration process for the
removal of methyl orange and methylene blue dyes, the
membranes exhibiting enhanced photocatalytic performance
under visible light (Athanasekou et al., 2014). GO-TiO2 com-
posite membranes were also prepared for exclusive use as
filtration membranes (i.e. without photocatalysis involved) to
remove hazardous dyes, such as methyl orange and rhoda-
mine B (Xu et al., 2013a).
In the present work, three photocatalysts, i.e. lab-made
bare TiO2 and GO-TiO2, as well as the reference material
in photocatalysis, Evonik Degussa (P25), were assembled on
flat sheet filtration cellulose membranes. The resulting
photocatalytic ultrafiltration membranes were tested for the
photodegradation and mineralization of a pharmaceutical
compound, diphenhydramine (DP), and an organic dye,
methyl orange (MO), under both near-UV/Vis and visible
light irradiation and in continuous operation mode. In
addition, an innovative freestanding GO membrane was
intercalated between the cellulose membrane and the
photocatalytic layer. Besides studies with the model pol-
lutants in distilled water (DW), the effect of the presence of
Cl
anions on the removal of the pollutants was also
studied in simulated brackish water (SBW) and seawater
(SSW).
2. Experimental
2.1. Chemicals and materials
Natural graphite (99.9995%), diphenhydramine hydrochloride
(C17H21NO$H2O, 99%), methyl orange (C14H14N3NaO3S, 99%),
ammonium hexafluorotitanate ((NH4)2TiF6, 99.99%) and boric
acid (H3BO3, 99%), were obtained from SigmaeAldrich. So-
dium chloride (NaCl, 99.5%) was supplied by Panreac. Mixed
cellulose ester (MCE) membranes with 0.45 mm pore size,
diameter of 4.7 cm and thickness of 140 mm were purchased
from Whatman™. The commercial TiO2 material (P25) was
supplied by Evonik Degussa GmbH (P25) and was used as a
reference material. It consists of 80% of anatase phase and
20% of rutile (manufacturer's data).
2.2. Synthesis of graphene oxide and GO-TiO2
Graphite oxide was obtained by the oxidative treatment of
commercial graphite (20 mm, SigmaeAldrich) following the
modified Hummers method, as described elsewhere
(Hummers and Offeman, 1958; Pastrana-Martı́nez et al., 2012).
The oxidized material was dispersed in water, sonicated for
1 h and centrifugated at 3000 r.p.m. to obtain a suspension of
graphene oxide (GO).
GO-TiO2 (hereafter referred as GOT) was prepared by liquid
phase deposition method according to our previous work
(Pastrana-Martı́nez et al., 2012). Briefly, (NH4)2TiF6 (0.1 mol L
1)
and H3BO3 (0.3 mol L
1) were added to a GO dispersion that
was then heated at 60  C for 2 h under vigorous stirring. Then
the resulting powder was treated as 200  C under N2 flow. The
carbon loading (~4.0 wt.%) was selected taking into account
the composite presenting the highest photocatalytic activity
under UV/Vis and visible light irradiation in our previous work
 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
(Pastrana-Martı́nez et al., 2012). Bare TiO2 was prepared using
the same methodology but without the addition of GO (here-
after referred as TiO2). The steps for the synthesis of GO, TiO2
and GO-TiO2 composite are shown in Scheme 1.
2.3. Preparation of photocatalytic membranes
GOT, TiO2 and P25 catalysts were selected to prepare the
corresponding photocatalytic membranes following a
methodology adapted from elsewhere (Morales-Torres
et al., 2014), in that case for the preparation of CNT
buckypapers over a polytetrafluoroethylene commercial
membrane. In a typical procedure, the photocatalyst was
dispersed in a 2 g L
1 aqueous solution during 10 min by
using an ultrasonic processor (UP400S, 24 kHz). The MCE
membrane was cut into a piece with 1.4 cm of diameter
and placed into a filtration device under vacuum. 5 mL of
the photocatalyst dispersion was added slowly (ca. 10 mg
of photocatalyst). When the filtration was completed, the
membrane was dried under N 2 flow for 10 min, a homo-
geneous deposition of the photocatalyst being obtained
with an effective area of ca. 1.54 cm 2 . The membranes
prepared with GOT, TiO 2 and P25 were labelled as M-GOT,
M
TiO 2 and M-P25, respectively.
The membrane presenting the highest photocatalytic ac-
tivity was also modified by intercalating a freestanding GO
membrane between the MCE membrane and the photo-
catalyst layer. First, 2 mL of GO dispersion (1.5 g L
1) was
filtered thought a MCE membrane, and a homogeneous GO
layer was obtained above the MCE membrane (labelled as M-
GO). Then, a uniform multi-layer membrane was prepared by
adding 5 mL of the GOT composite suspension (2 g L
1) on top
of the M-GO membrane (M-GO/GOT).
2.4. Characterization techniques
The morphology and roughness of the membranes were
analysed by scanning electron microscopy (SEM) using a FEI
Scheme 1 e Diagram for the synthesis of graphene oxide
(GO) by the modified Hummers' method (grey), as well as
for the preparation of TiO2 (white) and GO-TiO2 composite
(grey and white) by the liquid phase deposition method.
181
Quanta 400FEG ESEM/EDAX Genesis X4M instrument. The
membranes were frozen and broken under liquid nitrogen.
The microscope was equipped with a special multiple sample
holder, in which the broken membranes were vertically
positioned to analyse their cross-sections.
The overall porosity (ε) of the membranes was determined
by the gravimetric method (Cui et al., 2013). After measuring
the dry weight of the membranes, they were immersed in
distilled water overnight to assure solvent penetration into
the membrane pores. Then, their wet weight was registered.
The porosity was calculated by applying the following
equation:
ðmw
md Þ=rw
εð%Þ ¼ .  100 (1)
ðmw
md Þ=rw þ md rp
where mw and md are the weights of the wet and dry mem-
branes, respectively, and rw and rp are the water (0.997 g cm
3)
and polymer (1.520 g cm
3) densities, respectively (Maim et al.,
1947). Four different membranes were used to determine an
average value of the porosity, and the standard deviation
found was always lower than ±1%.
N2 adsorptionedesorption isotherms at
196  C were ob-
tained in a Quantachrome NOVA 4200e multi-station appa-
ratus. The apparent surface area (SBET) was determined by
applying the BrunauereEmmett-Teller (BET) equation
(Brunauer et al., 1938). The volume of N2 adsorbed at a relative
pressure of 0.95 (Vp) was obtained from the adsorption iso-
therms, which corresponds to the sum of the micro- and
mesopore volumes according to Gurvitch's rule (Rouquerol
et al., 1999). The BarretteJoynereHalenda (BJH) method
(Barrett et al., 1951) was applied to the desorption branch of
the N2 isotherms to determine the average mesopore diameter
(dpore).
The hydrophilicity of the membrane surface was deter-
mined by contact angle (q) measurements using an Attension
apparatus (model Theta) that allowed image acquisition and
data analysis. The measurements were performed at room
temperature (25  C) using the water drop method on dry
membranes. The contact angle was measured at least in 5
different locations to get the average value.
The optical properties of the samples were analysed by UV/
Vis diffuse reflectance spectroscopy using a JASCO V-560 UV/
Vis spectrophotometer, equipped with an integrating sphere
attachment (JASCO ISV-469), barium sulphate being used as
reference. The reflectance spectra were converted to equiva-
lent absorption Kubelka-Munk units by the instrument soft-
ware (JASCO). The band-gap was determined by plotting the
transformed Kubelka-Munk values as a function of the energy
of light.
The materials point of zero charge (pHPZC) was determined
following a pH drift test described elsewhere (Pastrana-
Martı́nez et al., 2013b). Briefly, solutions with varying initial
pH (2e12) were prepared using HCl (0.1 mol L
1) or NaOH
(0.1 mol L
1) and 50 mL of NaCl (0.01 mol L
1) as electrolyte.
Each solution was contacted with 0.15 g of the material and
the final pH was measured after 24 h of continuous stirring at
room temperature. The pHPZC of the material was determined
by intercepting the obtained final-pH vs. initial-pH curve with
the straight line final-pH ¼ initial-pH.
182 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
2.5. Evaluation of membranes performance
The photocatalytic activity and permeability of the mem-
branes were studied in dead-end filtration mode at ambient
temperature (25  C) and pressure under near-UV/Vis and
visible light irradiation, using a lab-scale set up represented in
Fig. 1. The system consists of a glass made cylindrical reactor
with the membrane attached by using a Viton® o-ring
(maintaining the effective area of ca. 1.54 cm2). A reservoir
containing the fresh pollutant solution was magnetically
stirred and continuously purged with air flow. DP
(3.40  10
5 mol L
1) and MO (3.05  10
5 mol L
1) were used as
model pollutants in distilled water (DW) at natural pH condi-
tions (pHDP ¼ 5.9 and pHMO ¼ 6.1). Experiments were also
performed using simulated brackish water (SBW, 0.5 g L
1 of
NaCl) and seawater (SSW, 35 g L
1 of NaCl). The pollutant
solution was continuously introduced in the reactor at a flow
rate of ca. 0.25 mL min
1 using a peristaltic pump. This flow
rate allows an adequate residence time into the reactor and
was selected from preliminary studies at different flow rates.
In a typical run, the solution was passed through the reactor
for a long time (up to 180 min) before turning on the lamp, to
saturate the membrane with the tested pollutants (ca. 60 min
needed) and to achieve the initial pollutant concentration in
the reactor outlet (i.e. C180 min in dark ¼ C0 min under radiation). In
the particular case of the M-GO/GOT membranes a total of
48 h dark period was needed before turning on the lamp.
The irradiation source consisted of a Heraeus TQ 150
medium-pressure mercury vapour lamp. A DURAN® glass
water cooling jacket was used to obtain irradiation in the
near-UV irradiation (l > 350 nm) and to control the operating
temperature. A cut-off long pass filter was used (l > 430 nm)
for visible light experiments. The photon flow entering the
reactor was ca. 33 mW cm
2 and 2.8 mW cm
2 under near-UV/
Vis and visible light irradiation, respectively, as determined by
integrating the irradiance spectra obtained by using a UV/Vis
spectroradiometer (USB2000þ, OceanOptics, USA). In the
photocatalytic experiments, two dark/bright cycles were car-
ried out in order to evaluate the anti-fouling properties of the
membranes. 2 mL of sample were systematically collected
from the exit of the photocatalytic membrane reactor (i.e.
before the permeate reservoir) at intervals of ca. 20 min, in
order to measure the pollutant concentration. Blank experi-
ments were also performed in the absence of catalyst in order
to characterize direct photolysis and the filtration capacity of
the commercial MCE membrane.
The concentration of DP was determined by HPLC with a
Hitachi Elite LaChrom system equipped with a Hydrosphere
Fig. 1 e Schematic diagram of the photocatalytic lab-scale set up.
C18 column. The concentration of MO was determined by
UVeVis spectrophotometry at 464 nm in a Jasco V-560 spec-
trophotometer. The total organic carbon (TOC) was deter-
mined for selected samples using a Shimadzu TOC-5000A
analyser. Chloride ions were monitored by ion chromatog-
raphy (Metrohm 881 Compact IC) using a Metrosep A Supp 7-
250 column.
The permeate flux (L m
2 h
1) was calculated according to
the following equation:
V
F¼ (2)
Axt
where V (L) is the volume of the solution permeated during the
experiment, A represents the effective membrane area (m2),
and t denotes time (h).
3. Results and discussion
3.1. Membranes characterization
The cross-section and top view of the prepared membranes
were analysed by SEM. The corresponding images of M
TiO2
and M-GOT at different magnifications are shown in
Figs. 2aef, respectively. The total thicknesses of the mem-
branes are gathered in Table 1.
Cross-sectional images of M
TiO2 and M-GOT membranes
(Figs. 2 aeb and dee, respectively) show that the corre-
sponding photocatalytic materials (top) were homogeneously
deposited on the MCE membrane (bottom) without appre-
ciable presence of cracks, holes or another defects, even if
considered that these membranes were fractured for SEM
analysis. Both membranes (M
TiO2 and M-GOT) showed
some differences in the packing and morphology of the
deposited material. M
TiO2 (Figs. 2aeb and Table 1) presented
a good distribution of TiO2 particles on the substrate mem-
brane with a uniform thickness of TiO2 around 35 mm, while
the overall thickness of the M
TiO2 membrane was ~175 mm
(ca. 140 mm for MCE only). The top view of the M
TiO2 mem-
brane (Fig. 2c) shows spherical particles forming larger ag-
gregates with anatase crystallites of around 4e5 nm in size
(Pastrana-Martı́nez et al., 2012).
The cross-sectional images of the M-GOT membrane
(Figs. 2dee) show a uniform GOT layer deposited on the MCE
membrane with a thickness of ~65 mm, almost twice as that
observed for TiO2 in the case of M-TiO2. The amount of pho-
tocatalyst deposited in both membranes was the same.
Therefore, the larger thickness observed for M-GOT can be
 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0 183

Fig. 2 e SEM micrographs at different magnifications of both M¡TiO2 (aeb: cross section; c: top view) and M-GOT (dee: cross
section; f: top view) membranes.

Table 1 e Characterization of membranes.

Sample Thickness (mm) SBET (m2 g
1) Vp (cm3 g
1) dpore (nm) pHPZC ε (%) q ( ) Eg (eV)
a
MCE 140 e e 450 e 74 e e
M-P25 179 65 0.13 29.8 6.3 66 18 3.2
M-TiO2 175 118 0.11 29.0 3.5 65 17 3.1
M-GOT 205 117 0.15 31.6 3.2 71 11 2.9
a
Data provided by the manufacturer.
184 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
attributed to the lower density of the GOT catalyst (resulting
from the arrangement of the GOT platelets, as shown in Fig. 2f
inset), which differs from that of TiO2 spherical particles. The
top view of the GOT composite (Fig. 2f) indicates a good TiO2
distribution on both sides of the GO sheets, i.e. a good self-
assembly of the TiO2 nanoparticles on GO (Pastrana-
Martı́nez et al., 2012).
Regarding the M-P25 membrane (images not shown), a
homogeneous layer of P25 with a thickness of ~39 mm (more
similar to M
TiO2 than to M-GOT) was observed over the MCE
membrane, with a typical morphology of joined P25
nanoparticles.
The total porosity (ε) of the membranes was determined by
the gravimetric method and the results are collected in Table
1. The porosity of the modified membranes (65e71%) was
lower than that determined for the MCE membrane (74%), but
that of M-GOT was the nearest to MCE (71%). The BET surface
area (SBET) values were comparable for M-GOT (117 m2 g
1) and
M
TiO2 (118 m2 g
1), and both are higher than that of M-P25
(65 m2 g
1). The total pore volume (Vp) was significantly higher
for M-GOT when compared to M
TiO2, evidencing that a
pronounced porosity is created when GO sheets and TiO2 are
in contact. The immobilization of the photocatalytic materials
on the MCE support (0.45 mm pore size) leads to a membrane
with lower pore size (29.8, 29.0 and 31.6 nm for M-P25, M
TiO2
and M-GOT, respectively). Therefore, the prepared mem-
branes could be considered in the range of ultrafiltration
(1 nme0.1 mm). The contact angles, also shown in Table 1,
were very low for all membranes (<18 ), indicating the high
hydrophilicity of the membranes' surface. In particular, M-
GOT presents the lowest contact angle (11 ), which can be of
relevance to treat high water fluxes and to avoid membrane
fouling (Abdelrasoul et al., 2013).
3.2. Removal of pollutants in dark phase
Fig. 3 shows DP and MO removals under dark conditions in
filtration experiments performed using both DW and SBW at
Fig. 3 e Pollutant removal obtained for the membranes by using filtration with DW and SBW in dark conditions during the
first 60 min; Blank ¼ MCE membrane.
natural pH and 25  C. The results are given for a period where
saturation of all membranes was observed (60 min), and
before starting to increase the concentration of pollutants in
the reactor outlet.
These results indicate that there is a low removal of the
pollutants (ca. 0.2%) when using the commercial MCE mem-
brane in the dark regardless the liquid media (i.e. DW or SBW,
respectively labelled as Blank_DW or Blank_SBW in Fig. 3).
Regarding the experiments with DW, DP removal was signif-
icantly higher for M-GOT (12%), M
TiO2 (9%) and M-P25 (3%)
than for MCE. The same tendency was observed for MO, i.e.
the highest MO removal was obtained with the M-GOT
membrane (8%), lower DP removals being obtained with the
other membranes, namely 6% with M
TiO2 and 5% with M-
P25. The removal of these pollutants can be related with two
different phenomena occurring under dark conditions,
adsorption or retention (Xu et al., 2013a). In order to discrim-
inate between both, some screening filtration experiments
were performed (not shown) suggesting that the pollutant
removal on the photocatalytic membranes may be attributed
to DP adsorption.
Taking into account the speciation diagrams of both DP
and MO (Figs. 4a and b, respectively), the pKa of DP and MO are
ca. 8.9 (Hein and Jeannot, 2001) and 4.2 (de Araujo et al., 2000),
respectively. Since the pH in experiments with DW was near
to 6.0 (5.9 for DP and 6.1 for MO), this means that DP is posi-
tively charged while MO is negatively charged under our
experimental conditions. In addition, the pHPZC of both bare
TiO2 and GOT materials are ca. 3.1 and 2.9, respectively (Table
1). Therefore, M
TiO2 and M-GOT membranes are expected to
present negatively charged surfaces at the pH of the experi-
ments (~6.0), explaining why the adsorption of DP is higher
(electrostatic interactions) than in that of MO (electrostatic
repulsions) when both membranes were employed. In the
case of M-P25, MO removal was higher than DP removal in DW
because the M-P25 membrane surface is practically un-
charged (pHPZC ca. 6.3) at the operating pH conditions (~6.0).
Overall, the lower adsorption capacity for the M-P25
 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
Fig. 4 e Speciation diagram of (a) DP and (b) MO in aqueous solutions at 25  C.
membrane compared to both M
TiO2 and M-GOT membranes
can be ascribed to SBET (Table 1) which is higher for bare-TiO2
and GOT (118 and 117 m2 g
1, respectively) than for P25
(65 m2 g
1).
The effect of the presence of NaCl (SBW) on the removal of
DP and MO under dark conditions is also shown in Fig. 3. The
results indicate that the presence of NaCl in water is not really
affecting the adsorption of these pollutants onto the mem-
branes, probably due to the low amount of Cl
ions in the SBW
prepared solutions (NaCl 0.5 g L
1). In addition, only ca. 4e5%
of the initial Cl
concentration was removed during the
filtration experiments under dark conditions (data not
shown). The Cl
ions are not retained in the membranes due
to their (i) low molecular size (167 pm for the Cl
ionic radius),
(ii) competitive adsorption with the pollutant molecules,
together with (iii) electrostatic repulsions in the case of M-GOT
and M-TiO2.
3.3. DP photocatalytic degradation
The efficiency of the membranes for the photocatalytic
degradation of DP under near-UV/Vis (closed symbols) and
visible light irradiation (open symbols) is shown in Figs. 5a
and b, respectively for DW and SBW, where C0 represents
the DP concentration in the dark after 180 min. It can be
observed that the MCE membrane (Blank, cross symbol) did
not have any photocatalytic performance by itself under
near-UV/Vis (and also under visible light irradiation, not
shown), using either DW or SBW, and that DP is a very re-
fractory pollutant when irradiated in the absence of a
catalyst, as previously reported (Pastrana-Martı́nez et al.,
2012). M-GOT and M-P25 are very active photocatalytic
membranes, presenting comparable efficiencies under near-
UV-Vis irradiation in terms of DP removal during the first
photocatalytic cycle (ca. 73% in DW and ca. 60% in SBW).
The lowest DP removal was obtained for M
TiO2 (43% in
DW and 39% in SBW). The same tendency was observed in
terms of TOC removal in 240 min, i.e. M-P25, M
TiO2 and M-
GOT produced TOC reductions of 35%, 18% and 35%,
respectively, in DW (Fig. 5c). For comparison purposes,
previous experiments carried out with the same irradiation
lamp and pollutants in slurry/batch mode (Pastrana-
Martı́nez et al., 2013a; 2014; 2012) were compared with
those obtained in the present work. Although suspended
185
photocatalysts have higher catalytic efficiency in compari-
son with the photocatalytic membranes, the separation of
the photocatalyst particles from the treated water is not
necessary when using the membranes, which is an impor-
tant technical advantage for real scale applications.
The photocatalytic DP degradation under visible light illu-
mination in DW (Fig. 5a) was significantly higher for M-GOT
(ca. 28%) than for TiO2 membranes (ca. 5% with both M
TiO2
and M-P25) indicating that TiO2 is not active under visible light
illumination, in agreement with previous reports where
similar materials were used in the form of powder slurries for
the photodegradation of different pollutants (Fotiou et al.,
2013; Maroga Mboula et al., 2013; Pastrana-Martı́nez et al.,
2012). This observation confirms that the addition of GO into
the TiO2 matrix increases light absorption in the visible
spectral range, resulting from a decrease in the band-gap
energy (2.9, 3.1, and 3.2 eV for M-GOT, M
TiO2 and M-P25,
respectively, Table 1). In our previous studies (Pastrana-
Martı́nez et al., 2012; 2013b), the high activity of the GOT
composite was attributed to the good TiO2 distribution in the
composite containing ca. 4% of GO content, leading to good
assembly and interfacial coupling between the GO sheets and
TiO2 nanoparticles (as can be observed in the SEM micrograph
of M-GOT, Fig. 2f).
Regarding the results with SBW (Fig. 5b), it can be
concluded that the presence of NaCl (0.5 g L
1) slightly de-
creases the photocatalytic efficiency for DP under near-UV/Vis
(i.e. 60%, 39% and 61% for M-P25, M
TiO2 and M-GOT,
respectively in the first cycle), in comparison with DW (Fig. 5a).
TOC removal (Fig. 5c) follows a trend similar to that observed
for DP degradation, i.e. lower mineralizations were also ob-
tained in the presence of NaCl (27%, 12% and 28% for M-P25,
M
TiO2 and M-GOT, respectively). This effect could be due to
the presence of Cl
ions acting as holes and hydroxyl radical
scavengers, affecting the efficiency of the photocatalytic pro-
cess. The presence of Cl
anions in SBW could generate less
reactive species such as chlorine radicals (Cl) and dichloride
anion radicals (Cl
2 ) (Eqs. (3)e(5)), which are less reactive than
HO radicals (De Laat and Le, 2006). The formation of these less
reactive species may decrease the efficiency of the process.
These results are in agreement with those published in liter-
ature (Sirtori et al., 2010; Yap and Lim, 2011), where a negative
effect of Cl
ions was observed for the photocatalytic degra-
dation of other pollutants.
186 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
Fig. 5 e Photocatalytic degradation of DP under near-UV/
Vis and visible light irradiation in continuous mode with
(a) DW and (b) SBW. (c) Total organic carbon (TOC) removal
in 240 min; Blank ¼ MCE membrane under near-UV/Vis.
Cl
þ hþ
vb #Cl
$
Cl
þ HO$ #ClOH$

Cl$ þ Cl
H Cl$

2 (5)
Membrane cleaning can be required after long operating
periods of time with salty water. Different physical, chemical
and biological cleaning methods can be applied depending on
the fouling nature (organic, inorganic, or biological), fouling
mechanism (adsorption, deposition or pore blockage) and
membrane properties (Shi et al., 2014; Wang et al., 2014).
3.4. MO photocatalytic degradation
The same type of photocatalytic experiments were performed
using MO as model pollutant. The photocatalytic activity of M-
P25, M
TiO2 and M-GOT membranes under both near-UV/Vis
and visible light irradiation are shown in Figs. 6a and b for DW
and SBW, respectively. It is observed that MO is poorly
removed when the commercial MCE membrane is used
(Blank). The M-GOT membrane showed significantly higher
photocatalytic activity for MO abatement than the other
membranes tested in DW (51% and 5% for M-P25, 39% and 4%
for M
TiO2, 65% and 19% for M-GOT, under near-UV/Vis and
visible light irradiation, respectively). The same trend was
(3)
(4)
Fig. 6 e Photocatalytic degradation of MO under near-UV/
Vis and visible light irradiation in continuous mode with
(a) DW and (b) SBW. (c) Total organic carbon (TOC) removal
in 240 min; Blank ¼ MCE membrane under near-UV/Vis.
 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
found concerning the respective TOC removal (Fig. 6c). The
photodegradation of MO from the first to the second cycle was
practically the same for all membranes when using DW as
solvent (Fig. 6a).
For the experiments with SBW (Fig. 6b), once again the
membrane prepared with the GOT composite exhibited the
highest photocatalytic activity under near-UV/Vis and visible
light irradiation (respectively, 52% and 13% reduction of the
initial MO concentration) and also the highest TOC removal
(respectively, 31% and 9%, Fig. 6c). However, the presence of
NaCl leads to a slight decrease in MO degradation, in com-
parison with DW, regardless of the membranes employed
(Fig. 6b). The same trend was observed in terms of TOC
removal (Fig. 6c). As previously observed for DP removal, MO
degradation decreases from the first to the second bright
cycle, reinforcing the idea that Cle ions are scavengers of holes
and hydroxyl radicals.
3.5. Permeate flux and anti-fouling of the photocatalytic
membranes
Regardless of the membrane employed, the photocatalytic
degradation of both pollutants tested was practically the same
from the first to the second bright cycle when using DW as
solvent (Figs. 5a and 6a for DP and MO, respectively). These
results indicate that all membranes are quite stable and that
fouling did not take place. This observation also suggests that
parent pollutants/intermediate compounds are not irrevers-
ibly accumulated on the membrane surface, which guaran-
tees that the membranes can maintain high permeate fluxes
for long operation periods (Zhang et al., 2014a).
The permeate flux was monitored for experiments per-
formed with DP. Figs. 7a and b show the results obtained for
DW and SBW, respectively, in 240 min. The flux performance
of the MCE membrane (blank) in darkness, UV/Vis and visible
light irradiation, did not change during the experiments. In
addition, the permeate flux of the lab-made membranes (M-
P25, M
TiO2 and M-GOT) was slightly lower in comparison
with that of the blank MCE membrane, probably due to the
nanostructured materials that were deposited over MCE and
that have influence on the membranes porosity.
The pollutants can stay adsorbed on the membrane sur-
face and can partially block the pores under dark conditions.
The permeate flux increased for M-P25, M
TiO2 and M-GOT
Fig. 7 e Permeate flux (measured in 240 min) of the membranes under dark conditions, near- UV/Vis and visible light
irradiation in DP experiments with (a) DW and (b) SBW.
187
membranes under near-UV/Vis irradiation, probably as a
consequence of the photocatalytic degradation of DP mole-
cules adsorbed on the membrane surface or to a higher hy-
drophilicity of the materials under near-UV/Vis irradiation.
However, a similar increase of the water flux was not always
observed under visible light illumination. Only for M-GOT the
permeate flux was higher at such conditions than in the
darkness, which can be related to the high photocatalytic
activity of this membrane under visible illumination, when
compared with the others (Fig. 7), and/or to its higher hydro-
philicity (contact angles of 11 , 17 and 18 for M-GOT, M
TiO2
and M-P25, respectively). Therefore, overall, M-GOT performs
better than M
TiO2 and M-P25.
3.6. Hierarchically ordered M-GO/GOT membrane
A free-standing GO membrane (M-GO) was introduced be-
tween the commercial MCE membrane and the GOT layer
with the aim to improve the performance of the M-GOT
membrane. The resulting membrane was referred as M-GO/
GOT and its efficiency was evaluated for the degradation of
DP, first in darkness and then under visible and near-UV/Vis
irradiation. The experiments were carried out using DW and
simulated seawater (SSW e 35 g L
1 NaCl).
The cross-sectional SEM micrographs of the prepared M-
GO/GOT membrane are shown in Fig. 8. Uniform GO layers (ca.
6.5 mm of total thickness) between the MCE membrane and the
GOT photocatalyst are observed. The GO membrane consists
of several GO sheets (inset of Fig. 8b: M-GO) and the GOT
catalyst was homogeneously deposited on top of M-GO. The
M-GO/GOT membrane has an overall thickness of ~200 mm, in
which ~50e55 mm are due to the GOT photocatalyst. The
thickness of the GOT layer in M-GO/GOT seems to be smaller
than in the case of M-GOT (65 mm), both prepared with the
same amount of photocatalyst. This smaller thickness could
be due to a stronger packing of the GOT particles during the
preparation step, since a longer time of filtration was required
to prepare M-GO/GOT.
Figs. 9a and b show the results obtained in terms of DP
removal and permeate flux, respectively. The M-GO/GOT
membrane exhibited a significantly higher DP removal under
dark conditions in DW when compared to M-GOT (i.e, 12 and
42% for M-GOT and M-GO/GOT, respectively). The DP con-
centration increased after the initial fast DP removal and
188 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0

Fig. 8 e Cross-sectional M-GO/GOT SEM micrographs at different magnifications. (a) 1000 x and (b) 2500 x (inset corresponds
to the freestanding GO membrane: M-GO).

before turning on the lamp, but a long dark period (up to 48 h)
was needed for the saturation of the M-GO-GOT membrane
with the DP pollutant while only 180 min where required in
the case of the M-GOT membrane.
These results indicate that the adsorption and/or exclusion
of pollutants of the M-GOT membrane was increased by the
introduction of the free-standing GO membrane, the M-GO/
GOT membrane presenting a higher DP adsorption capacity in
the filtration process. As expected, the removal capacity was
lower when the experiments were performed in SSW (10%)
due to the high content of Cl
anions (35 g L
1) that could
reduce the electrostatic interactions between catalysts and
pollutants due to a screening effect (Stuart et al., 1991), i.e. the
Cl
anions might pair with the positive DP molecules and
reduce the electrostatic interactions with the catalyst surface,
and/or could block the active sites of the catalyst (Kamble
et al., 2008).
Regarding photocatalytic activity, the M-GO/GOT mem-
brane also showed a relatively good activity for DP degrada-
tion under near-UV/Vis and visible light irradiation; however,
the M-GOT membrane performed better (Figs. 5a and b
respectively). Since the thickness of the GOT layer in M-GO/
GOT was smaller (~50e55 mm) than in the case of M-GOT
(65 mm), one possible explanation is that the photocatalyst is
more compacted in the case of M-GO/GOT and, therefore, less
exposed to the irradiation, an important aspect when dealing
with photocatalytic reactors. The longer time of filtration
required by the presence of GO is the main synthesis limita-
tion regarding this issue. The permeate flux increased for the
M-GOT membrane under near-UV/Vis and visible light irra-
diation (Fig. 7), probably as a consequence of the photo-
catalytic degradation of DP molecules adsorbed on the
membrane surface, or to a higher hydrophilicity of the ma-
terials under irradiation. The permeate flux was also moni-
tored for the M-GO/GOT membrane in darkness and under
visible and near-UV/Vis irradiation (Fig. 9b). As expected, the
results showed a decrease of the permeate flux under dark
conditions. However, an increase in permeate flux was sub-
sequently observed under visible light and near-UV/Vis irra-
diation, as in the case of M-GOT. This effect was more

Fig. 9 e (a) DP removal and (b) variation of the permeate flux in DW and SSW (35 g L¡1) with the M-GO/GOT membrane under
dark conditions, near-UV/Vis and visible light irradiation.
 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
significant for the experiments in DW, and can be attributed to
the photocatalytic activity, hydrophilicity and anti-fouling
properties of the prepared membranes.
4. Conclusions
Different photocatalytic membranes were synthesized with P25
(M-P25), bare-TiO2 (M
TiO2) and GO-TiO2 composite. The pre-
pared membranes were compared in terms of photocatalytic
activity using distilled water and simulated brackish water and
seawater. The photocatalyts were homogeneously deposited
without appreciable presence of cracks, holes or another de-
fects. All the membranes presented high activity and stability in
consecutive light-dark cycles under continuous mode. During
the photocatalytic reaction, the permeate flux increased due to
the high hydrophilicity of the membranes and larger contami-
nant removal by photodegradation. The presence of NaCl had a
substantially detrimental effect on the photocatalytic perfor-
mance which may be due to the chloride anions acting as hole
and hydroxyl radical scavengers. In general, the M-GOT mem-
brane showed significantly higher photocatalytic activity for the
pollutants abatement than the other membranes tested. In
addition, an innovative hierarchically membrane referred as M-
GO/GOT was prepared by intercalating a GO membrane in the
structure of the M-GOT. This membrane showed higher
pollutant removal under dark conditions and good performance
under visible and near-UV/Vis irradiations. However, the M-GOT
membrane performed better, probably due to the higher
compactness in the case of M-GO/GOT as a consequence of the
synthesis conditions required for its preparation.
Acknowledgements
Financial support for this work was provided by project
NORTE-07-0202-FEDER-038900 (NEPCAT), financed by FEDER
(Fundo Europeu de Desenvolvimento Regional) through ON2
(Programa Operacional do Norte) and QREN, and by project
PTDC/AAC-AMB/122312/2010, co-financed by FCT (Fundaça ~o
para a Cie^ncia e a Tecnologia) and FEDER through Programme
COMPETE (FCOMP-01-0124-FEDER-019503). This work was
partially co-financed by FCT/MEC and FEDER under Programe
PT2020 (Project UID/EQU/50020/2013), and by QREN, ON2 and
FEDER through project NORTE-07-0124-FEDER-000015. LMPM
and SMT acknowledge financial support from FCT grants
SFRH/BPD/88964/2012 and SFRH/BPD/74239/2010, respec-
tively. AMTS acknowledges the FCT Investigator 2013 Pro-
gramme (IF/01501/2013), with financing from the European
Social Fund and the Human Potential Operational Programme.
Technical assistance by Dr. Carlos Sa
and CEMUP team with
SEM analysis is gratefully acknowledged.
references
Abdelrasoul, A., Doan, H., Lohi, A., 2013. In: Nakajima, H. (Ed.),
Mass Transfer - Advances in Sustainable Energy and
Environment Oriented Numerical Modeling. InTech.
189
Albu, S.P., Ghicov, A., Macak, J.M., Hahn, R., Schmuki, P., 2007.
Self-organized, free-standing TiO2 nanotube membrane for
flow-through photocatalytic applications. Nano Lett. 7 (5),
1286e1289.
Amalraj Appavoo, I., Hu, J., Huang, Y., Li, S.F.Y., Ong, S.L., 2014.
Response surface modeling of Carbamazepine (CBZ) removal
by Graphene-P25 nanocomposites/UVA process using central
composite design. Water Res. 57, 270e279.
Athanasekou, C.P., Morales-Torres, S., Likodimos, V.,
Romanos, G.E., Pastrana-Martinez, L.M., Falaras, P.,
Dionysiou, D.D., Faria, J.L., Figueiredo, J.L., Silva, A.M.T.,
2014. Prototype composite membranes of partially reduced
graphene oxide/TiO2 for photocatalytic ultrafiltration water
treatment under visible light. Appl. Catal. B 158e159,
361e372.
Baek, Y., Kim, C., Seo, D.K., Kim, T., Lee, J.S., Kim, Y.H., Ahn, K.H.,
Bae, S.S., Lee, S.C., Lim, J., Lee, K., Yoon, J., 2014. High performance
and antifouling vertically aligned carbon nanotube membrane
for water purification. J. Membr. Sci. 460, 171e177.
Barrett, E.P., Joyner, L.G., Halenda, P.P., 1951. The Determination
of pore volume and area distributions in porous substances. I.
Computations from nitrogen isotherms. J. Am. Chem. Soc. 73
(1), 373e380.
Brunauer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in
multimolecular layers. J. Am. Chem. Soc. 60 (2), 309e319.
Cui, Z., Hassankiadeh, N.T., Lee, S.Y., Lee, J.M., Woo, K.T.,
Sanguineti, A., Arcella, V., Lee, Y.M., Drioli, E., 2013.
Poly(vinylidene fluoride) membrane preparation with an
environmental diluent via thermally induced phase
separation. J. Membr. Sci. 444, 223e236.
Chen, X., Mao, S.S., 2007. Titanium dioxide nanomaterials:
synthesis, properties, modifications, and applications. Chem.
Rev. 107 (7), 2891e2959.
Das, R., Ali, M.E., Hamid, S.B.A., Ramakrishna, S.,
Chowdhury, Z.Z., 2014. Carbon nanotube membranes for
water purification: a bright future in water desalination.
Desalination 336, 97e109.
de Araujo, R.E., Gomes, A.S.L., de Araújo, C.B., 2000.
Measurements of pKa of organic molecules using third-order
nonlinear optics. Chem. Phys. Lett. 330 (3e4), 347e353.
De Laat, J., Le, T.G., 2006. Effects of chloride ions on the iron(III)-
catalyzed decomposition of hydrogen peroxide and on the
efficiency of the Fenton-like oxidation process. Appl. Catal. B
66 (1e2), 137e146.
Dreyer, D.R., Park, S., Bielawski, C.W., Ruoff, R.S., 2010. The
chemistry of graphene oxide. Chem. Soc. Rev. 39 (1), 228e240.
Fan, W., Lai, Q., Zhang, Q., Wang, Y., 2011. Nanocomposites of
TiO2 and reduced graphene oxide as efficient photocatalysts
for hydrogen evolution. J. Phys. Chem. C 115 (21), 10694e10701.
Fotiou, T., Triantis, T.M., Kaloudis, T., Pastrana-Martı́nez, L.M.,
Likodimos, V., Falaras, P., Silva, A.M.T., Hiskia, A., 2013.
Photocatalytic degradation of microcystin-LR and off-odor
compounds in water under UV-a and solar light with a
nanostructured photocatalyst based on reduced graphene
oxideeTiO2 composite. Identification of intermediate
products. Ind. Eng. Chem. Res. 52 (39), 13991e14000.
Gao, P., Liu, Z., Tai, M., Sun, D.D., Ng, W., 2013. Multifunctional
graphene oxideeTiO2 microsphere hierarchical membrane for
clean water production. Appl. Catal. B 138e139, 17e25.
Gao, Y., Hu, M., Mi, B., 2014. Membrane surface modification with
TiO2egraphene oxide for enhanced photocatalytic
performance. J. Membr. Sci. 455, 349e356.
Hein, J., Jeannot, M., 2001. Drug distribution: a guided-inquiry
laboratory experiment in coupled homogeneous and
heterogeneous equilibria. J. Chem. Educ. 78 (2), 224.
Huang, X., Wang, L., Zhou, J., Gao, N., 2014. Photocatalytic
decomposition of bromate ion by the UV/P25-Graphene
processes. Water Res. 57, 1e7.
190 w a t e r r e s e a r c h 7 7 ( 2 0 1 5 ) 1 7 9 e1 9 0
Hummers, W.S., Offeman, R.E., 1958. Preparation of graphitic
oxide. J. Am. Chem. Soc. 80 (6), 1339e1339.
Kamble, S.P., Mangrulkar, P.A., Bansiwal, A.K., Rayalu, S.S., 2008.
Adsorption of phenol and o-chlorophenol on surface altered
fly ash based molecular sieves. Chem. Eng. J. 138 (1e3), 73e83.
Kim, J., Van der Bruggen, B., 2010. The use of nanoparticles in
polymeric and ceramic membrane structures: review of
manufacturing procedures and performance improvement for
water treatment. Environ. Pollut. 158 (7), 2335e2349.
Long, M., Qin, Y., Chen, C., Guo, X., Tan, B., Cai, W., 2013. Origin of
visible light photoactivity of reduced graphene oxide/TiO2 by
in Situ hydrothermal growth of undergrown TiO2 with
graphene oxide. J. Phys. Chem. C 117 (32), 16734e16741.
Maim, C.J., Genung, L.B., Fleckenstein, J.V., 1947. Densities of
cellulose esters. Ind. Eng. Chem. Res. 39 (11), 1499e1504.
Maroga Mboula, V., He
quet, V., Andre s, Y., Pastrana-
Martı́nez, L.M., Don ~ a-Rodrı́guez, J.M., Silva, A.M.T., Falaras, P.,
2013. Photocatalytic degradation of endocrine disruptor
compounds under simulated solar light. Water Res. 47 (12),
3997e4005.
Morales-Torres, S., Silva, T.L.S., Pastrana-Martinez, L.M.,
Brandao, A.T.S.C., Figueiredo, J.L., Silva, A.M.T., 2014.
Modification of the surface chemistry of single- and multi-
walled carbon nanotubes by HNO3 and H2SO4 hydrothermal
oxidation for application in direct contact membrane
distillation. Phys. Chem. Chem. Phys. 16 (24), 12237e12250.
Pan, J.H., Zhang, X., Du, A.J., Sun, D.D., Leckie, J.O., 2008. Self-
etching reconstruction of hierarchically mesoporous f-TiO2
hollow microspherical photocatalyst for concurrent
membrane water purifications. J. Am. Chem. Soc. 130 (34),
11256e11257.
Pastrana-Martı́nez, L.M., Morales-Torres, S., Kontos, A.G.,
Moustakas, N.G., Faria, J.L., Don ~ a-Rodrı́guez, J.M., Falaras, P.,
Silva, A.M.T., 2013a. TiO2, surface modified TiO2 and graphene
oxide-TiO2 photocatalysts for degradation of water pollutants
under near-UV/Vis and visible light. Chem. Eng. J. 224, 17e23.
Pastrana-Martı́nez, L.M., Morales-Torres, S., Likodimos, V.,
Falaras, P., Figueiredo, J.L., Faria, J.L., Silva, A.M.T., 2014. Role
of oxygen functionalities on the synthesis of
photocatalytically active grapheneeTiO2 composites. Appl.
Catal. B 158e159, 329e340.
Pastrana-Martı́nez, L.M., Morales-Torres, S., Likodimos, V.,
Figueiredo, J.L., Faria, J.L., Falaras, P., Silva, A.M.T., 2012.
Advanced nanostructured photocatalysts based on reduced
graphene oxideeTiO2 composites for degradation of
diphenhydramine pharmaceutical and methyl orange dye.
Appl. Catal. B 123e124, 241e256.
Pastrana-Martı́nez, L.M., Morales-Torres, S., Papageorgiou, S.K.,
Katsaros, F.K., Romanos, G.E., Figueiredo, J.L., Faria, J.L.,
Falaras, P., Silva, A.M.T., 2013b. Photocatalytic behaviour of
nanocarboneTiO2 composites and immobilization into hollow
fibres. Appl. Catal. B 142e143, 101e111.
Pendergast, M.M., Hoek, E.M.V., 2011. A review of water treatment
membrane nanotechnologies. Energy Environ. Sci. 4 (6),
1946e1971.
Peters, T., 2010. Membrane technology for water treatment.
Chem. Eng. Technol. 33 (8), 1233e1240.
Rouquerol, F., Rouquerol, J., Sing, K.S.W., 1999. Adsorption by
Powders and Porous Solids. Principles, Methodology and
Applications. Academic Press, London.
Shannon, M.A., Bohn, P.W., Elimelech, M., Georgiadis, J.G.,
Marinas, B.J., Mayes, A.M., 2008. Science and technology for
water purification in the coming decades. Nature 452 (7185),
301e310.
Shi, X., Tal, G., Hankins, N.P., Gitis, V., 2014. Fouling and cleaning
of ultrafiltration membranes: a review. J. Water Process Eng. 1,
121e138.
Sirtori, C., Agüera, A., Gernjak, W., Malato, S., 2010. Effect of water-
matrix composition on trimethoprim solar photodegradation
kinetics and pathways. Water Res. 44 (9), 2735e2744.
Stuart, M.A.C., Fleer, G.J., Lyklema, J., Norde, W.,
Scheutjens, J.M.H.M., 1991. Adsorption of ions, polyelectrolytes
and proteins. Adv. Colloid Interface Sci. 34, 477e535.
Tu, W., Zhou, Y., Zou, Z., 2013. Versatile graphene-promoting
photocatalytic performance of semiconductors: basic
principles, synthesis, solar energy conversion, and
environmental applications. Adv. Funct. Mater. 23 (40),
4996e5008.
Ulbricht, M., 2006. Advanced functional polymer membranes.
Polymer 47 (7), 2217e2262.
Wang, Z., Ma, J., Tang, C.Y., Kimura, K., Wang, Q., Han, X., 2014.
Membrane cleaning in membrane bioreactors: a review. J.
Membr. Sci. 468, 276e307.
Xu, C., Cui, A., Xu, Y., Fu, X., 2013a. Graphene oxideeTiO2
composite filtration membranes and their potential
application for water purification. Carbon 62, 465e471.
Xu, J., Chang, C.Y., Hou, J., Gao, C., 2013b. Comparison of
approaches to minimize fouling of a UF ceramic membrane in
filtration of seawater. Chem. Eng. J. 223, 722e728.
Yap, P.S., Lim, T.T., 2011. Effect of aqueous matrix species on
synergistic removal of bisphenol-A under solar irradiation
using nitrogen-doped TiO2/AC composite. Appl. Catal. B 101
(3e4), 709e717.
Zhang, H., Quan, X., Chen, S., Zhao, H., 2006. Fabrication and
Characterization of silica/titania nanotubes composite
membrane with photocatalytic capability. Environ. Sci.
Technol. 40 (19), 6104e6109.
Zhang, X., Wang, D.K., Diniz da Costa, J.C., 2014a. Recent
progresses on fabrication of photocatalytic membranes for
water treatment. Catal. Today 230, 47e54.
Zhang, X., Wang, D.K., Lopez, D.R.S., Diniz da Costa, J.C., 2014b.
Fabrication of nanostructured TiO2 hollow fiber photocatalytic
membrane and application for wastewater treatment. Chem.
Eng. J. 236, 314e322.